Pizzolato, M (2023) AM-PM facial image comparison for forensic human identification. Other thesis, Liverpool John Moores University.

Preview

Text 2023pizzolatomphil.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (7MB) | Preview

Abstract

The uniqueness of the face has been studied concerning its application for human identification. In particular, morphological facial comparison has been used to verify the identities of living individuals in forensic investigations and has received considerable attention, including validation studies and the development of international guidelines (FISWG, 2019a; Bacci et al., 2021b). At the same time, the application of morphological comparison in post-mortem identification has been recognised (Olivieri et al., 2018; Cappella et al., 2021), but remains understudied and require further validation. In light of this, this study investigates the reliability of morphological facial comparison using an unconstrained sample of ante-mortem (AM) and post-mortem (PM) images of recent deceased (N=29), and aims to develop a methodological protocol combining the structure of the living individual's guidelines and accounts for the early post-mortem changes that affect the face. The reliability of the morphological method for PM identification is investigated by performing the comparisons and documenting the process using the protocol, and by testing it with three observers on selected AM-PM pairs (N=15).
The key findings suggest that AM-PM photographic comparison using the proposed protocol could help to narrow down the potential matches of a PM subject. Moreover, in some cases, it is also possible to provide a single correct AM target for a PM subject with a high level of confidence. While the validity of the protocol requires further investigation beyond this study, its design, which includes the addition of a stage for analysis of the decomposition changes affecting the PM subjects and a holistic preliminary analysis of AM-PM pairs to exclude the obvious non-matches, seems advantageous. The interobserver study results indicate that both absolute agreement (ICC=0.813) and agreement on the level of support (Kendal’s K=0.885) are satisfactory. Overall, the limitation of the study includes the small interobserver random sample (15 pairs), the absence of non-matching AM subjects in the sample pool tasks, and presence of a very limited number of PM subjects showing signs of post-mortem changes (e.g. facial bloating).
This study could benefit forensic casework, in particular in the Disaster Victim identification (DVI) procedures, when other AM primary identifiers are scarce. Future studies should investigate the application of PM photographic facial comparison using a larger database simulating an open disaster scenario, while also further validating and testing the applicability of the developed method with different observers.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	post mortem
Subjects:	H Social Sciences > H Social Sciences (General)
Divisions:	Biological & Environmental Sciences (from Sep 19)
SWORD Depositor:	A Symplectic
Date Deposited:	25 Jan 2024 12:18
Last Modified:	25 Jan 2024 12:19
DOI or ID number:	10.24377/LJMU.t.00022205
Supervisors:	Wilkinson, C, Liu, C and Cattaneo, C
URI:	https://researchonline.ljmu.ac.uk/id/eprint/22205

View Item